NYC Rolling Sales Market Analysis

Personal Project

Project Overview

Guided by my philosophy that "data is not about numbers, it's about people," I approached this project as an Architect of Clarity. I recognized that nonsensical entries, like the frequent $0 sale prices (often representing deed transfers between family members), could significantly skew market averages if not properly contextualized. My strategy was to move beyond surface-level cleaning and apply a "Logic-First Audit" to preserve the dataset's numerical integrity while respecting the real-world transitions it represents.

Objective

To establish a "source of truth" by identifying and resolving critical structural inconsistencies. The goal was to cleanse invalid entries—such as zero-value placeholders and missing building data—to ensure accurate calculation of average property metrics and reliable predictive modeling.

Business Analytics

Strengths

Extensive scope covering over 22,000 property records across all 5 NYC Boroughs, providing a high-volume sample for meaningful analysis

Weakness

Raw data initially contained significant "noise," including nonsensical $0 sales and missing build dates that required manual logic rules to resolve.

Opprotunities

A cleansed dataset allows organizations to spot emerging market trends and build models that accurately predict future sale values.

Threats

Structural inconsistencies in property identifiers (like the BBL keys) can lead to flawed market models and poor investment decisions if the audit phase is skipped.

Solution

I implemented a multi-layered cleansing solution involving 4 major logic audits:

Structural Standardization: Uniformly formatted inconsistent address entries and removed non-numerical text from square footage columns to protect mathematical integrity.
Placeholder Resolution: Located and corrected invalid entries, such as "0" for ZIP CODE, and replaced non-cash transaction indicators (-) with blanks to prevent them from being treated as market-value sales.
Gap Identification: Flagged critical missing values in YEAR BUILT and SALE PRICE columns to ensure they did not pollute averages in future analysis.
Trend Visualization: Designed insight-driven charts to translate complex borough-wide data into actionable trends.

22,000+ Rows Validated: Systematically audited and standardized property records borough-wide.
Borough-Wide Coverage: Successfully integrated data from Manhattan, Bronx, Brooklyn, Queens, and Staten Island.
Foundation for Modeling: Established the structural work required for deep analysis of property metrics and advanced market modeling.

Results

Technical Skills Used

Data Auditing & Cleansing, Structural Consistency Auditing, Data Visualization, Pattern Recognition

Tools Used

Microsoft Excel